Read-only tree can't read head (Re: Web interface update)

[Martin Pool]

On Sat, 2005-05-21 at 20:04 -0400, Aaron Bentley wrote:

> > Yeah, it should just ignore failures to write the statcache.  I think
> > for trees opened readonly it's still useful to try to update the cache
> > so that it can accelerate access where possible.
> 
> That means a separate lock for the statcache, I think.

No, I don't think it's needed because the cache is just a single file
and it's renamed into place.

-- 
Martin